J . Org. Chem. 1985, 50, 474-479
474
Response of Triflates to Solvent Ionizing Power. A YOTfScale Based on 7-Norbornyl Triflate Xavier Creary* and Steven R. McDonald Department of Chemistry, University of Notre Dame, Notre Dame, Indiana 46556
Received June 26. 1984 A solvent effect study on rates of solvolysis of 7-norbornyl triflate (1) showed a relatively small overall response values was observed despite the involvement to solvent ionizing power parameters. A poor correlation with YOT~ of a cationic intermediate. Triflate 1, which is very resistant to SN2-type reactions, has been used to define a new set of solvent ionizing power values, Y m . The triflate derivative of pivaloin (2), 1-methylcyclopropyl triflate (3), and 6,6-(ethylenedioxy)-endo-bicyclo[2.2.l]hept-2-~~1 triflate (5) all gave poor rate correlations with Yon values, despite the involvement of cationic intermediates in solvolysis. Greatly improved correlations with YoTrvalues were observed. endo-2-Methyl-exo-bicyclo[3.1.0]hex-2-yl triflate (4) gave an mom value which was more indicative of a cationic intermediate than the mOTs value. Previously reported data for a-phenyltrifluoroethyl triflate (13) and ( E ) -and (2)-butenyl triflates 17 and 18 showed improved correlations when rates were plotted against YoTf values. These studies all suggest that the solvation requirements of the developing, relatively stable trifluoromethanesulfonate anion are somewhat different from tosylate or chloride ion. Low rate responses to increasing solvent polarity in triflate solvolyses are not inconsistent with involvement of cationic intermediates.
One of the primary mechanistic criteria employed to establish the intermediacy of carbocationic intermediates in solvolysis processes is a large increase in rate with increasing solvent polarity. Winstein and Grunwald’ expressed this rate increase quantitatively. A linear free energy relationship (eq 1) was developed where k is the log ( k / k o ) = mY
(1)
solvolysis rate a t a given substrate in a given solvent, lzo is the rate in the standard solvent, 80% aqueous ethanol, Y is a solvent parameter called the solvent ionizing power, and m is a substrate parameter which gives the response of a substrate to changes in solvent ionizing power. The substrate used by Winstein and Grunwald to define Y values was tert-butyl chloride. Over the years, variations of the original Winstein-Grunwald equation have appeared. The importance of the nucleophilic character of a solvent in determining solvolysis rates of certain substrates was recognized quite early. This led to the introduction of a four-parameter equation (eq 2)2 which atlog ( k / k o ) = 1N
+ mY
(2)
tempts to express rates as a function of solvent ionizing power and solvent nucleophilicity N . Another variation used by Schleyer, Bentley, and Schadt3 is a three-parameter equation (eq 3) which contains the adjustable paramlog ( k / k o ) = (1- 8) 1% ( ~ / ~ o ) c H ~ + o TQ~ log
(k/kO)P-AdOTs
(3)
eter Q. This equation allows one to determine to what extend a substrate’s reactivity resembles methyl tosylate (Q= 0) or 2-adamantyl tosylate (Q = 1). For a completely limiting case (a k , process), eq 3 reduces to eq 4 which has log ( h / k 0 ) = mYoTs
(4)
the same form as the original Winstein4runwald equation where Q and m are equivalent. Equation 4 uses a set of Y values which is based on the defining substrate, 2(1) (a) Grunwald, E.; Winstein, S. J. Am. Chem. SOC.1948, 70, 846-854. (b) Winstein, S.; Grunwald, E.; Jones, H. W. Ibid. 1951, 73, 2700-2707.
adamantyl t o ~ y l a t e .This ~ secondary substrate was used to define Y values since it solvolyzes by a k , mechanism; i.e., there is negligable nucleophilic solvent assistance in a range of solvent^.^ This simple equation has retained popularity and has recently been used to correlate a number of substrates solvolyzing by cationic proce~ses.”~ Over the years, we have been interested in solvolytic reactions of substrates containing the triflate leaving group. We reportedea in a limited study (two solvents) that cyclopropyl triflates gave a fairly small rate response to increasing solvent ionizing power in an aqueous acetone solvent. Schiavelli and Stangg have also observed decreased solvent responses in solvolyses of certain vinyl triflates in aqueous ethanol systems. We also foundebthat the triflate derivative of pivaloin gave a relatively small rate response. However, we were confident, based on the rearrangements observed, that these substrates were solvolyzing via cationic intermediates. It therefore appears that the m value criterion fails to confirm the presence of cationic intermediates for certain triflates when Y values based on tert-butyl chloride and 2-adamantyl tosylate are employed. We have therefore carried out a detailed examination of the response of triflates 1-5 to changing solvent ionizing power. These substrates all solvolyze, giving cationic intermediates of widely differing structure. How do these substrates response? Answers to this question are reported herein. Solvolysis of 7-Norbornyl Triflate (1). The rate behavior of 7-norbornyl triflate (1) has been examined in a variety of solvents. This substrate reacts via a cationic intermediate, 6, which has been suggested by recent studies (4) (a) Fry, J. L.; Lancelot, C. J.; Lam, L. K. M.; Harris, J. M.; Bingham, R. C.; Raber, D. J.; Hall, R. E.; Schleyer, P. v. R. J. Am. Chem. SOC. 1970, 92, 2538-2540. (b) Fry, J. L.; Harris, J. M.; Bingham, R. C.; Schleyer, P. v. R. Ibid. 1970, 92, 2540-2542. (c) Schleyer, P. v. R.; Fry, J. L.; Lam, L. K. M.; Lancelot, C. J. Zbid. 1970,92, 2542-2544. (5) (a) Allen, A. D.; Jansen, M. P.; Koshy, K. M.; Mangru, N. N.; Tidwell, T. T. J . Am. Chem. SOC. 1982,104, 207-211. (b) Allen, A. D.; Ambridge, I. C.; Che, C.; Michael, H.; Muir, R. J.; Tidwell, T. T. Ibid 1983, 105, 2343-2350. (6) (a) Creary, X.; Geiger, C. C. J. Am. Chem. SOC.1982, 104, 4151-4162. (b) Creary, X.; Geiger, C. C.; Hilton, K. Ibid. 1983, 105, 2851-2858. (c) Creary, X. Ibid. 1984, 106, 5568. (7) (a) Bentley, T. W.; Carter, G. E. J. Org. Chem. 1983, 48, 579-584. (b) Brown, H. C.; Ravindranathan, M.; Chloupek, F. J.; Rothberg, I. J. Am. Chem. SOC.1978 100, 3143-3149. (8) (a) Creary, X. J. Am. Chem. SOC.1976,98,6608-6613. (b) Creary, X.; Geiger, C. C. J. Am. Chem. SOC.1983, 105, 7123-7129. (9) Schiavelli, M. D.; Jung, D. M.; Vaden, A. K.; Stang,P. J.; Fisk, T.; Morrison, D. S. J . Org.Chem. 1981, 46, 92-95.
0022-3263/85/1950-0474$01.50/00 1985 American Chemical Society
Triflate Response to Solvent Ionizing Power
-1
2 -
J. 0%.Chem., Vol. 50, No. 4, 1985 475
3 -
t d s - *.
.,."
L I P , , I . -
2
.
i
0
I
3
2
5
4
yo,.
1 -
-6
of Sunko'O to be bent. Rate data are given in Table I and presented graphically as a function of YoTa in Figure 1. While there are generally rate increases with solvent ioni z i i power, there is a relatively poor correlation (r = 0.92) with Ym (or Yt.BuCI) values over the entire range of solvents studied. The overall m value (0.58) is also rather small. While a linear relationship is observed in aqueous ethanol solvents, rates are faster in these solvents than those expected on the basis of Yme values relative to other solvents." The rate in trifluoroacetic acid is also considerably slower than that expected on the basis of its Ym value. Triflate 1 has been chosen as a model k, substrate for triflate solvolyses since it reacts a t convenient rates in a variety of solvents.12 Additionally, as suggested for 2adamantyl systems, the solvolysis of 1 is not susceptible to nucleophilic solvent assistance. This has been verified by the previous ob~ervation'~ that 7-norbornyl tosylate does not give SN2-typedisplacement of tosylate even when treated with methoxide or cyanide in methanol at 200 O C . Additionally our studies show that potassium thiophenoxide reacts with 7-norbornyl triilate less readily than with cyclohexyl mesylate despite the fact that the triflate leaving group is a substantially better leaving group (approximately lo5times more reactive)Iobthan the mesylate nucleofuge. Examination of molecular models shows that the reluctance of triflate 1 to give SN2-typereactions is due in part to the C2C3hydrogens of 1which hinder the preferred approach of a nucleophile. Since nucleophilic solvent H,,OTi
Figure 1. Plot of log k for solvolysis of 1 vs. YoTs.
~
-
2
-
-
1
0
1
-
_
-
1
2
,
-
/
3
_
1
I
4
5
"07,
Figure 2. Plot of log k for solvolysis of 2 vs. YoT,. OTI
I
0
I - B u -CH-C-!-B"
-2
-,
1 0 YOT?
Figure 3. Plot of log k for solvolysis of 2 vs. YoTI.
(10)Cation 6 is the structure suggested by MINDO/3 d d a t i o n s . A aomewhat different bent cation is suggested by MIND0/2 calculations. See: Sunko, D.E.; Vaneik, H.; Deljac, V.; Milun, M. J. Am. Chem. Soe. 1983,105, 5364-5368. For previous studies on 7-norbomyl triflate, ~ e e ref 8. See also: (b) Su, T. A,; Sliwinski, W. F.; Sehleyer, P. Y. R. J. Am. Chem. Soe. 1969,91,53865388. For studies on 7-norbomyl towlate, see: (c) Winstein, S.; Shatavaky, M.; Norton, C.; Woodwsld, R. B. J. Am. Chem. Sac. 1955, 77,4183. (d) Gassman, P. G.; Hornback, J. M. Ibid. 1967.89.2487-2488. (e) Gassman, P. G.; Hornback, J. M.; Marshall, J. L. Ibid. 1968,90,623&6239. (0Miles, F. B. Ibid. 1967.89, 2488-2489. ( 9 ) Miles. F. B. Ibid. 1968, 90,1265-1628. (11) There me instances where aqueous ethanol data gives linear mY correlations and other aqueou solvents give different (but linear) correlations. This type of dispersion is often observed when Y values based on tert-butyl chloride are used. Ideally one would not see such dispersions if completely analogous mechanisms operated. For typical examples, gee ref 2. (12) 2-Adamantyl triflatd waa not chosen as a model triflate since it has extremely high reactivity in more highly ionizing solvents. (13) Gassman, P. G.; Hornback, J. M.: Paseone, J. M. Tetrahedron Lett. 1971, 1415-1416.
assistance in solvolysis of 1 is presumably negligible the rate data in Table I has been used to define a Ym ... scale. Values are given in Table 11. Solvent Effects on Reactivitv of Triflates 2-5. The reactivity of triflates 2-5 has been examined in a variety of solvents. Triilate 2 solvoyhs by a k, route as evidenced 5 0 CH3 0 CHrC-CH-C-c-Bu I I I1
by the observation of exclusively rearranged p r o d ~ c t s . 8 ~ ~ ' ~
476
J. Org. Chem., Vol. 50, No. 4, 1985
Creary and McDonald Table I. Solvolysis Rates of Triflates
compound
XT
solvent a EtOH
HOAc
80% EtOH
60% EtOH TFE HCO,H 97% H F l P TFA
HOAc TFE HCO,H 97% H F l P TFA
T, “C
compound
k , s-’
114.9 95.0 25.0b
1.46 X 1.96 x 10-5 2.01 x 1 0 . ~
125.0 100.0 25.0 25.0bsC 95.0 75.0 25.0b 80.0 60.0 25.0b 80.0 60.0 25.0b 60.0 40.0 25.0 60.0 35.0 25.0b 25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0
5.25 x 10-4 3.81 x 10-5 1.07 X 10.’ 1.37 x 10-9 4.89 X 10 5.46 x 10-4 6.41 X 6.50 x 10-4 6.32 x 10.~ 4.88 X 10.’ 1.38 X 1.53 x 10-5 1.62 x 10-7 2.08 x 1 0 . ~ 1.57 x 10-5 1.80X 1.19 x 9.10 x 2.89 X 3.72 X 5.98 X 10.‘ 4.15 x 1 0 . ~ 1.28 x 10-4 4.15 X 4.36 x 10-5 2.22 x 10-4 1.82 x 10-4 2.0 x 10-4
solventa
T, “C
EtOH HOAc 80% EtOH 60% EtOH TFE HCO,H 97% H F l P TFA EtOH 80% EtOH 60% EtOH HOAc TFE HCO,H 97% H F l P
25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0
4.03 x 10-5 5.52 x 10-5 6.53 x 10-4 2.71 x 10-3 9.24 x 10-4 5.07 x 10-3 6.41 x 10-3 5.9 x 10-3 4.44 x 10-5 6.42 x 1 0 - 4 2.86 x 10-3 5.82 x 10.~ 5.47 x 10-3 1.14 X l o - ’ 7.11 X lo-’
EtOH HOAc 80% EtOH 60% EtOH TFE HCO,H 97% H F l P
25.0 25.0 25.0 25.0 25.0 25.0 25.0
1.00 x 7.53 x 8.05 x 2.21 x 1.29 x 4.18 x 4.51 x
k , s‘l
10-4 10-5 10-4 10-3 10-3 10-3 10-3
a EtOH contains 0.025 M 2,6-lutidine; 80% EtOH contains 20% (by volume) water; 60% EtOH contains 40% ( b y volume) water; HOAc contains 1%acetic anhydride and 0.1 or 0.05 M NaOAc; TFE, 0.025 M 2,6-lutidine in trifluoroethanol; HC0,H contains 0.05 M sodium formate; 97% H F l P contains 3% (by weight) water in hexafluoroisopropyl alcohol and 0.05 M 2,g-lutidine; TFA contains 0.5% trifluoroacetic anhydride and 0.2 M sodium trifluoroacetate. Extrapolated value. Reference l o b .
Table 11. Solvent Ionizing Power Values solvent YOTf YOTs EtOH -1.50 -1.75 80% EtOH 0.00 0.00 60% EtOH 0.88 0.92 HOAc -1.78 -0.61 TFE 0.40 1.80 HCOzH 1.49 3.04 97% HFlP 1.65 3.61 TFA 1.76 4.57 Table 111. Correlation of Triflate Solvolysis Rates with Y Values cmpd ( p ; Syz; nln mOTf (r; spy;nla 1 0.58 (0.921; 0.50; 8) 1.00 (definition) 2 0.27 (0.869; 0.32; 8) 0.49 (0.990; 0.09; 8) 3 0.36 (0.905; 0.35; 8) 0.64 (0.992; 0.10; 8) 4 0.59 (0.969; 0.27; 7) 0.85 (0.966; 0.29; 7) 5 0.34 (0.903; 0.29; 7) 0.54 (0.998; 0.05; 7) 13 0.46 (0.830; 0.63; 8) 0.84 (0.961; 0.31; 8) 13 (omit TFA) 0.60 (0.908; 0.50; 7) 0.94 (0.993; 0.14; 7) 17 0.27 (0.688; 0.52; 5) 0.59 (0.969; 0.18; 5) 18 0.08 (0.171; 0.88; 5) 0.51 (0.671; 0.66; 5) O r
= correlation coefficient; Syz = standard error of estimate; n
= number of data points.
Despite the intermediacy of cation 7, the entire rate spread in the solvents examined is only a factor of 53. A poor correlation (P = 0.869) is observed (Figure 2 and Table 111) when rates are plotted vs. YoTs.However, deviations are, in a qualitative sense, similar to those seen for 7-norbornyl (14) Creary, X. J. Org. Chem. 1979, 44, 3938-3945.
triflate. When rate data are plotted vs. Y o ~values f (Figure 3), triflate 2 gave a fairly good correlation (r = 0.990). The m value of 0.49 is still substantially lower than that for the defining substrate, 7-norbornyl triflate. This suggests that the “small” solvent response in 2 is due not only to the triflate leaving group but also to some other factor. This other factor may well be charge delocalization in the transition state for solvolysis of this k , substrate. Triflate 3 solvolyzes by a concerted ionization-ringopening mechanism, giving the allyl cation 10 as the initial intermediate.15 As in the case of 2, a poor correlation is ‘OTf
+‘
obtained when rates are plotted vs. YmSvalues (Table 111). When plotted against YoTf values (Figure 4), a much better correlation (r = 0.992) is obtained. The mOTfvalue of 0.64 is substantially less than 1.0, and, as in the case of 2, may be due to transition-state charge delocalization or less than “normaln transition-state charge development. Triflate 4 solvolyzes by a stepwise ionization to an unopened cyclopropyl cation 11 (a k , process) which then rapidly opens to an allylic cation 12.15*pb The exo orientation of the triflate group in this ring system precludes the concerted ionization-ring-opening process.16 The high (15) (a) Creary, X.J. Org. Chem. 1976,41,3734-3739; (b) 3740-3743. For studies on the corresponding tosylate, see: (c) Schleyer, P. v. R.; Sliwinski, W. F.; Van Dine, G. W.; Schollkopf,U.; Paust, J.; Fellenberger, K. J. A m . Chem. SOC.1972, 94, 125-133.
J. Org. Chem., Vol. 50, No. 4, 1985 477
Triflate Response to Solvent Ionizing Power
4 -
11 -
12 -
reactivity of 4 in trifluoroacetic acid, which generally gives the largest deviation in YmS plots, prevented measurement of rates in this solvent. The mms value obtained from the other solvents was 0.59 and the correlation was fair. A rate plot vs. YoTf values does not significantly change the correlation. However, the m m value is 0.85, a value more indicative of a k, process. Triflate 5, which is expected to be substantially less reactive than endo-2-norbornyl triflate due to the inductive effect of the ketal functionality,” showed trends similar to those of 2 and 3. The poor correlation seen with YoTs was substantially improved when rates were plotted vs. YoTf (Figure 5). For unknown reasons, the mOTf value (0.54) remains substantially smaller than unity. Correlation of Triflates 13 and 17 with Y m Values. TidwellSbhas done a detailed study of the reactivity of triflate 13 and concluded that this substrate solvolyzes via a cationic process, generating the “electron deficient cation” 14. A puzzling feature was the lack of correlation of rate I
Ph-CH-CF3
HoS
0
-2
14 -
0 Ts
I
CF3
+
HOS
P h- C
I
H3
-2
-
-3
-
-
5 -
1
-
0 -
- C F3
-
t I
H3
,
I
-2
I
0
-1
2
1
‘OTl
-
15
2
1
Figure 4. Plot of log k for solvolysis of 3 vs. YOTI.
-4
Ph-CI
0
-I
‘OTl
m
13 -
-
-3
0 Tf Ph-CH-CF3
3 3
1
16
with Ym values. This contrasted with the behavior of the tertiary analogue 15,5which gave a good correlation with Ym values. We have now plotted rate data for 13 vs. YoTf values (Figure 6), and a significantly improved correlation (r = 0.961) is seen. The trifluoroacetic acid data still deviates from the line defined by the other solvents, and omission of this data gave a correlation coefficient of 0.993 and an mOTfvalue of 0.94. Clearly, the behavior of 13 is more analogous to 7-norbornyl triflate than to 2-adamantyl tosylate and supports the suggested intermediacy of cation
Figure 5. Plot of log k for solvolysis of 5 vs. Yon. OTI
I
Ph-CH - C
o m i t TFA d a t a
F3
13 -
a TFA
14.
Attention was next turned to vinyl triflates 17 and 18. It had been previously notedla that these triflates gave a poor correlation with Yt-BuCIvalues, with rates in acetic acid, TFE, and TFA showing substantial deviations (lower reactivity) from the line defined by aqueous ethanol solvents. Our attempt (Table 111) to correlate the date for triflates 17 and 18 with YoTf values gave only limited H,
/OTf /C
=c,
CH3
CH3
17 -
,O T f
CH3, H
IC =c\ CH3
18 -
(16) (a) DePuy, C. H.; Schnack, L. C.; Hawser, J. W.; Wiedmann, W. J. Am. Chem. SOC.1965,87,4006. (b) Cristol, S.J.; Sequeira, R. M.; De Puy, C. H. Zbid. 1967, 4007-4008. (c) SchBllkopf, U.; Fellenberger, K.; Patach, M.; Schleyer, P. v. R.; Su, T.;Van Dine, G . W. Tetrahedron Lett. 1967,3639-3644. (17) For a study showing the deactivating inductive effect of the ketal functionality on solvolysis, see: Creary, X.; Rollin, A. J. J. Org. Chem. 1977,42, 4231-4238. (18) Summerville, R. H.; Senkler, C. A.; Schleyer, P. v. R.; Dueber, T. E.; Stang, P. J. J.Am. Chem. SOC.1974, 96,1100-1110.
-- 54 / ; / . / ,
,
,
1
2
~
,
-6
-
3
-
2
-
1
0
3
4
‘OTI
Figure 6. Plot of log k for solvolysis of 13 vs. YOTf. The solid line excludes data in TFA. success. With use of the five solvents for which data are available, an improved correlation is seen for 17 (mom = 0.59; r = 0.969), while a poor correlation is seen for 18. Rates of solvolysis of 18 in TFE and TFA still fall substantially below the line defied by acetic acid and aqueous ethanol solvents. The reason for the poor correlation for 18 with YoW values is uncertain but may be indicative of the importance of solvent nucleophilicity and/or basicity in solvolysis of 18. Summary and Conclusions In general, triflates solvolyze with less than “normal”rate increases with increasing solvent polarity. This overall
478 J . Org. Chem., Vol. 50, No. 4 , 1985
Creary and McDonald
t r e n d is t h o u g h t to be partially d u e t o t h e decreased solvation demands of the developing trifluoromethanesulfonate anion which should require less stabilization b y hydrogen bonding than chloride or p-toluenesulfonate anions. Over a large range of solvents, poor correlations of r a t e data with YoTsvalues were observed. Rates i n trifluoroacetic acid were considerably slower than expected o n the basis of YoTsvalues and rates i n ethanol-water mixtures were somewhat faster than expected. The decreased rates of triflate solvolysis i n trifluoroacetic acid m a y be reflecting reduced electrophilic catalysis in loss of the triflate group d u e to the low basicity of the trifluoromethanesulfonate moiety. 7-Norbornyl triflate has been used to define a set of solvent ionizing power values, Y O T p In general, triflate solvolysis rates gave superior correlations with these YoTf values than with YoTBor Y,.,, values. T h i s also suggests that the solvation requirements of the developing trifluoromethanesulfonate anion are somewhat different from tosylate or chloride. Lack of large rate increases with solvent polarity in solvolysis of triflates therefore does not rule o u t t h e involvement of cationic intermediates.
Experimental Section Triflates 1,19 3,15aand 4,15a were prepared as previously described. P r e p a r a t i o n of 6,6-(Ethylenedioxy)-exo-bicyclo[2.2.1]heptan-2-01. The previous procedurez0for preparation of this alcohol involved reaction of lithium aluminum hydride in N ethylmorpholine with exo-2,3-epoxy-6,6-(ethylenedioxy)bicyclo[2.2.l]heptane. The present procedure uses lithium triethylborohydride (Aldrich Chemical Co.) in di-n-butyl ether to open the epoxy ketal. Three-hundred milliliters of 1.0 M lithium triethylborohydride in tetrahydrofuran was added to 30 g of exo-2,3-epoxy-6,6(ethylenedioxy)bicyclo[2.2.1]heptanein a 1-L flask under nitrogen. The tetrahydrofuran was removed by evacuation of the flask using a water aspirator until a viscous oil remained. The atmosphere in the evacuated flask was replaced with nitrogen, and 200 mL of dry butyl ether was added. The mixture was heated and the temperature in the flask was maintained a t 120-125 "C for 10 h. The mixture was cooled to room temperature and about 200 mL of 10% aqueous sodium hydroxide was carefully added. The mixture was transferred to a separatory funnel and extracted with two portions of ether. The combined ether extracts were washed with water and saturated sodium chloride solution and dried over Na2S04. The solvent was removed by rotary evaporator and the residue was distilled to give 22.5 g (74%) of 6,6-(ethylenedioxy)-exo-bicyclo[2.2.1]heptan-2-ol,20 bp 73-75 "C (0.03 mm). Preparation of 6,6-(Ethylenedioxy)bicyclo[2.2.1]heptan2-one. Dimethyl sulfoxide (6.3 mL) was dissolved in 50 mL of methylene chloride and the mixture was cooled to -70 "C. Trifluoroacetic anhydride (9.3 g) was added slowly dropwise to the mixture at -70 "C. The hydroxy ketal obtained above (5.0 g) was dissolved in 5 mL of Me2S0 and 5 mL of CH2C12,and the solution was added dropwise to the -70 "C solution. The mixture was then stirred at -70 "C for 30 min and 12.3 mL of triethylamine was added. The mixture was allowed to warm to room temperature and ether was added, and the organic phase was extracted with three small portions of water. The organic phase was washed -with saturated NaCl solution and dried over Na2S04. The solvent was removed on a steam bath, and the residue was distilled to give 4.7 g of 6,6-(ethylenedioxy)bicyclo[2.2,l]heptan-2-one:bp 75-83 "C (0.16 mm); IR uc=o 1750 cm-'; NMR (CC14)6 4.2-3.7 (4 H, m), 2.73 (1H, m), 2.48 (1H, m), 2.35-1.55 (6 H, m). Anal. Calcd for CgH& C, 64.27; H, 7.19. Found: C, 64.34; H, 7.23. Preparation of 6,6-(Ethylenedioxy)-endo -bicyclo[2.2.1]heptan-2-01, A solution of 1.31 g of 6,6-(ethylenedioxy)bicyclo(19) Creary, X.; Benage, 2887-2891.
B.; Hilton, K. J . Org. Chem. 1983,
48,
(20) Meinwald, J.; Cadoff, B. C. J. Org. Chem. 1962, 27, 1539-1541.
[2.2.l]heptan-2-one in 8 mL of ether was added dropwise to a mixture of 0.26 g of LiAlH4 in 30 mL of ether. After being stirred for 30 min at room temperature, a solution of 0.10 g of NaOH in 0.9 g of water was carefully added dropwise. The mixture was stirred for 8 h and MgS0, was added. After filtration, the solvent was removed on a rotary evaporator. The entire product was chromatographed on 30 g of silica gel and eluted with increasing amounts of ether in Skelly solve F. The product eluted with 20-25% ether. After solvent remwal on a rotary evaporator, the residue was distilled to give 0.94 g (71%) of 6,6-(ethylenedioxy)-enclo-bicyclo[2.2,l]heptan-2-01: bp 90-93 "C (1.9 mm); NMR (CDCl,) d 4.4-3.5 (6 H, m), 2.4-0.8 (8 H, M). Anal. Calcd for CSH14Os: C, 63.51; H, 8.29. Found: C, 63.86; H 8.36. Preparation of Triflate 5. A solution of 0.50 g of the hydroxy ketal obtained above in 5 mL of CHPClzwas cooled to -50 "C and 0.38 g of 2,6-lutidine was added. T o the stirred solution under nitrogen was added 0.92 g of triflic anhydride, and the mixture was then allowed to warm to 0 "C. The mixture was then taken up into ether, washed rapidly with cold water, dilute HCl solution, and saturated NaCl solution, and dried over MgSO,. After filtration, the solvent was removed on a rotary evaporator, leaving 0.86 (97%) of crude triflate 5. Triflate 5 decomposes readily on standing in the absence of solvent a t room temperature. Even a t -20 "C, decomposition is substantial after 1 day. A CDCIB solution also slowly decomposes: NMR (CDCl,) 6 5.21 (1 H, m), 4.2-3.6 (4 H, m), 2.8-1.2 (8 H, m). Reaction of 7-Norbornyl Triflate with Thiophenoxide in Isopropyl Alcohol. A solution of potassium thiophenoxide in isopropyl alcohol (0.45 M) was prepared under nitrogen by dissolving 1.76 g of potassium metal in isopropyl alcohol, adding 5.0 g of thiophenol, and diluting to 100 mL. A solution of 0.428 g of triflate 1 in 7.8 mL of this solution was sealed under nitrogen in four tubes. After 23 h at room temperature, the contents of the first tube were taken up into ether and water, washed with dilute KOH solution and saturated NaCl solution, and dried over MgS04. Gas chromatography showed a trace of 7-norbornyl phenyl sulfide. After solvent removal on a rotary evaporator, the residue was analyzed by 300-MHz NMR which showed 96% unreacted 7-norbornyl triflate and 4% 7-norbornyl phenyl sulfide. An analogous reaction of cyclohexyl mesylate for 23 h a t room temperature gave 14% unreacted cyclohexyl mesylate and 86% cyclohexyl phenyl sulfide. The second tube was heated at 70 "C for 290 min and a similar workup followed. Analysis by 300-MHz NMR showed 22% unreacted 7-norbornyl triflate and 78% 7-norbornyl phenyl sulfide. Under the same conditions, cyclohexyl mesylate was completely consumed. The remaining two tubes were heated at 100 "C for 5 h and a similar workup followed. Gas chromatography showed no unreacted 1. After solvent removal on a rotary evaporator, 0.177 g (99%) of crude 7-norbornyl phenyl sulfide was obtained: NMR (CDC13) 6 7.7-7.0 ( 5 H, m) 3.23 (1 H, br s), 2.5-1.0 (10 H, m). Kinetics Procedures. Rates of solvolysis of triflates 1-5 in acetic acid, ethanol, trifluoroethanol, formic acid, and 97% hexafluoroisopropanol were determined by using the titrimetric method previously describeda6 Sealed ampules were not used for runs at 25 "C. For rapid rates, samples were quenched in acetic acid at about 16 "C, cooled in an ice bath, and titrated as rapidly as possible as previously described. Reactions of 1-4 in 80% ethanol and 60% ethanol were monitored by quenching 2-mL aliquots in 4 mL of absolute ethanol and titrating the liberated triflic acid with 0.01 M Et3N in ethanol. Reactions of 5 in 80% ethanol and 60% ethanol were buffered with 0.025M Et3N, and the unreacted Et3N was titrated with 0.01M HC104 in ethanol. Solvolysis of 1 in trifluoroacetic acid was monitored by 300-MHz NMR by following the decrease in intensity of the C-7 hydrogen at 6 3.97. The corresponding hydrogen in the product appears 0.07 ppm upfield in TFA. The reaction was monitored for 70 h and the NMR tube was maintained a t 25 "C between measurements. Solvolysis of 3 in TFA was monitored by quenching samples in cold CDCl, and following the appearance of the olefinic methylene hydrogens or the methylene hydrogens on the carbon bonded to the trifluoroacetoxy group of the product by 300-MHz NMR. Analyses were carried out rapidly since reaction continues to occur (but a t a greatly decreased rate) after quenching in CDCl3
J. Org.Chem. 1985, 50, 479-484 Acknowledgment is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, and to the National Science Foundation (CHE-8305820)for support of this research. We also acknowledge the NIH for a grant used to purchase the 300-MHz NMR spectrometer. Note Added in Proof Professor D. N. Kevill has examined 2-adamantyltriflate and has developed a Ym scale based on this substrate. We thank Professor Kevill for
479
informing us of his results prior to publication. Registry No. 1, 25354-43-2; 2, 71341-17-8; 3, 60153-74-4; 4, 60153-71-1; 5 , 94110-61-9; 13, 84877-46-3; 17, 24576-95-2; 18, 24577-69-3; exo-2,3-epoxy-5,5-(ethylenedioxy)bicyclo[2.2.l]heptane, 61557-99-1;6,6-(ethylenedioxy)-exo-bicyclo[2.2.1]heptan-2-ol, 61524-55-8; 6,6-(ethylenedioxy)bicyclo[2.2.l]heptan-2-one, 67594-65-4; 6,6-(ethylenedioxy)-endo-bicyclo[2.2.1]heptan-2-01, 94160-56-2; potassium thiophenoxide, 3111-52-2; 7-norbornyl phenyl sulfide, 94110-62-0; triflic anhydride, 358-23-6.
A Novel Dissociative Mechanism in Acyl Group Transfer from Aryl 4-Hydroxybenzoates in Aqueous Solventst Georgio Cevasco,t Giuseppe Guanti,* Andrew R. Hopkins,s Sergio Thea,**and Andrew Williams*$ Zstituto di Chimica Organica dell'Universit6, C.N.R., Centro d i Studio sui Diariloidi e loro Applicazioni, Corso Europa, Genova, Italy, and University Chemical Laboratory, Canterbury, England Received J u l y 17, 1984
The hydrolysis of aryl 4-hydroxybenzoate esters exhibits the kinetic rate law kOM = (k, + kb[OH-])/(l + [H+]/K,) where k b is the second-order rate constant for hydroxide ion attack on the ionized ester and K , is the ionization constant. The apparent second-order rate constant for the hydrolysis of the 2,4-dinitrophenyl ester (k,K,/K,) is some 340-fold larger than that determined from the Hammett correlation for the alkaline hydrolysis of substituted 2,4-dinitrophenyl benzoates known to possess a BA,2 mechanism. A slightly positive entropy of activation for k, and aniline trapping experiments for the 2,4-dinitrophenyl ester are consistent with a mechanism where a p-oxo ketene intermediate takes the reaction flux.
products
+
Correlation of kd(,/K, with the pK of the leaving phenol fits the equation k&,/K, = 10(-'.33pK+9.57)10(4,35PK+3.51) which favors the B A , mechanism ~ for poor leaving groups and the ElcB pathway for good leaving groups. There is no reason to postulate a borderline concerted displacement process in this case.
Introduction We are interested in the factors causing change in mechanisms of acyl group transfer in aqueous solution from associative paths such as the BA,2 process to dissociative mechanisms for example ElcB or SNL1 Whereas the S N 1 route for acyl group transfer (eq 1) is rare in RCOX
- -X-
fY-
RCO+ RCOY (1) I aqueous solution because the acylium ion I is unstable, the pathway can be favored if electron-donating features are included in the structure of R which could be RN-F RC-H,3 or O-.4 The acylium ion is thus effectively neutralized as isocyanate (II), ketene (III), or carbon dioxide (IV), respectively. One can conceive of many other ways in which R6-0
RtH&O
I RN=C=O
II
-O=&
I
1
t RCH=C=O
III
t O=C=O
IV V
the SN1intermediate could be stabilized to yield a species 'Dedicated to Professor Giuseppe Leandri on the occasion of his 70th birthday. Instituto di Chimica Organica dell'Universit6. *University Chemical Laboratory.
not normally thought of as an acylium ion.' We investigate here the possibility that the p-oxo ketene V is an ElcB intermediate. Species similar to V have been discussed in the literature and there is matrix isolation trapping evidence that the 0-oxo ketene has been i ~ o l a t e d . ~A nitrogen analogue (an o-imino ketene) has been observed in the mass spectrometer6 and similar species have been postulated to explain the formation of novel reaction products.' Ketenes and thioketenes fused into aromatic rings have been reported.8 The species are related to quinones and quinone methides which are well-known compounds; they are liable to be much more reactive than these species however because nucleophilic attack at the (1) (a) Preliminary accounts of some of this work have appeared: Thea, S.; Guanti, G.; Petrillo, G.; Hopkins, A. R.; Williams, A. J. Chem. SOC.,Chem. Commun. 1982, 577. Cevasco, G.; Guanti, G.; Thea, S.; Williams, A. Ibid. 1984,783. (b) Williams, A.; Douglas, K. T. Chem. Rev. 1975, 75, 627. (2) (a) Bender, M. L.; Homer, R. B. J . Org. Chem. 1965,30,3975. (b) Williams, A. J . Chem. Soc., Perkin Trans. IZ 1972, 808. (3) Bruice, T. C.; Holmquist, T. C. J.Am. Chem. Soc. 1968,90, 7136. (4) Sauers, C. K.; Jencks, W. P.; Groh, S. J. Am. Chem. SOC.1975,97, 5546. (5) Chapman, 0. L.; McIntosh,C. L.; Pacansky, J.; Calder, G. V.; Orr, G. J . Am. Chem. SOC.1973,95, 4061. (6) Williams, A.; Salvadori, G. J. Chem. SOC.B. 1971, 1105. (7) (a) Schultz, R.; Schweig, A. Tetrahedron Lett. 1979, 59. (b) Dennis, N.; Katritzky, A. R.; Parton, s. K. J. Chem. SOC.,Perkin Trans. I 1974, 750. (c) Dvorak, W.; Kolc, J.; Michl, J. Tetrahedron Lett. 1972, 3443. (d) Horspool, W. M.; Khandelwal, G. D. J . Chem. SOC.,Chem. Commun. 1970,257; J. Chem. SOC.C 1971,3328. ( e ) Torres, M.; Clement, A,; Bertie, J. E.; Gunning, H. E.; Strausz, 0. P. J. Org. Chem. 1978, 43, 2490. (8) (a) Segbold, G.; Heibl, C. Chem. Ber. 1977,110, 1225. (b) Alborz, M.; Douglas, K. T. J. Chem. Soc., Perkin Trans. 2 1982, 331.
0022-3263/85/1950-0479$01.50/00 1985 American Chemical Society
